The general enthusiasm for floating exchange rates in the 1970s was not shared by Continental European countries who, concerned about their internal market, experimented instead with regional schemes for exchange rate management. The EMS has proven quite successful in preventing large shifts in competitiveness between its member countries. Some observers have attributed this success to the presence of controls on capital movements; with the liberalization of these controls, further development of the EMS will only be feasible, it is argued, if member countries are prepared to forgo independent monetary policies.

In Discussion Paper No. 299, Programme Director Marcus Miller and Research Fellow Paul Weller examine the compatibility of exchange rate management and monetary autonomy, using a version of the Dornbusch model of exchange rate determination, with free trade and perfect capital mobility. Miller and Weller explore the effects of adjustable exchange rate bands in this model, using an approach recently introduced by Krugman. The authors analyse the implications for monetary policy and exchange rates of announcing rules for realigning the exchange rate bands when the rate hits the edge. The exchange rate is treated as a forward-looking asset price, whose value is independent of its past behaviour: arbitrage ensures that the current level of the exchange rate is its long-run equilibrium value discounted by expected cross-country interest rate differ- entials.
Miller and Weller consider the case where, when the exchange rate hits the edges of the band, market participants are unsure whether the money supply will be adjusted to eliminate the interest rate differential, i.e. the band is defended, or whether there will be a parity realignment with a simultaneous adjustment in the money supply. They analyse a variety of policy rules, including what they describe as `the complete monetary accommodation' rule: if the rate hits the edge of the band, the money stock target is adjusted exactly to accommodate the divergence of the price level from its equilibrium, and the centre of the exchange rate band is moved by the same percentage amount. Market participants are assumed to attach some given probability that the exchange rate band will be defended, i.e. the credibility of the rule. If the band is ever defended, the authorities establish full credibility; if a realignment does occur, the market does not change its perception of the probability of future realignments.
These assumptions allow Miller and Weller to describe the behaviour of the exchange rate as a function of the price level and to show how its path depends on the degree of credibility of the realignment rule. There exists a critical value for this credibility which exactly counteracts `bias in the band' and places the exchange rate on the same path as it would follow under free floating within the band.
If there is uncertainty about whether the realignment rule will be followed, the exchange rate always makes a discrete jump at the edge of the band. If the authorities establish full credibility for the original band by mounting a defence of the currency, the jump is inside the band, while if the authorities choose to realign, the currency jumps outside the band to an equilibrium corresponding to the new money stock target.
Of particular interest is the implication of a completely accommodating realignment rule which is fully credible. In this case the market is certain that when the rate hits the top (or bottom) edge there will be an upward (or downward) realignment: the band will never be defended. Intuition might suggest that the money stock and the exchange rate should follow a random walk, keeping the exchange rate at its purchasing power parity (PPP). In Miller and Weller's model, however, this is not the case: the exchange rate follows a curvilinear path about the locus of PPP. These deviations from PPP arise because the realignment rule introduces thresholds for monetary accommodation. The money stock remains unchanged within the band, and jumps to a new target level when there is a realignment: monetary policy is non- accommodating inside the prespecified thresholds. In addition, because monetary adjustment is discontinuous, the behaviour of the exchange rate betrays an element of `hysteresis': its relationship to the price level is not unique, but depends on the previous shocks hitting the system.
Miller and Weller's analysis shows that it is possible to combine some measure of monetary autonomy and some management of the exchange rate, while preserving free trade in both goods and financial assets. This constitutes an alternative to Obstfeld's `speculative attack' approach, which suggests that abolishing capital controls within the EMS will lead to spiralling devaluations

Exchange Rate Bands and Realignments in a Stationary Stochastic Setting
Marcus Miller and Paul Weller

Discussion Paper No. 299, March 1989 (IM)